Mathematical aspects of outer-product asynchronous content-addressable memories (Q911512): Difference between revisions

From MaRDI portal
Added link to MaRDI item.
ReferenceBot (talk | contribs)
Changed an Item
 
(One intermediate revision by one other user not shown)
Property / MaRDI profile type
 
Property / MaRDI profile type: MaRDI publication profile / rank
 
Normal rank
Property / cites work
 
Property / cites work: Stability and attractivity in associative memory networks / rank
 
Normal rank
Property / cites work
 
Property / cites work: Neural networks and physical systems with emergent collective computational abilities. / rank
 
Normal rank
Property / cites work
 
Property / cites work: Content-addressable memories. 2nd ed / rank
 
Normal rank
Property / cites work
 
Property / cites work: The existence of persistent states in the brain / rank
 
Normal rank
Property / cites work
 
Property / cites work: Analytic study of the memory storage capacity of a neural network / rank
 
Normal rank
Property / cites work
 
Property / cites work: A logical calculus of the ideas immanent in nervous activity / rank
 
Normal rank
Property / cites work
 
Property / cites work: On associative memory / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q3290963 / rank
 
Normal rank
Property / cites work
 
Property / cites work: The capacity of the Hopfield associative memory / rank
 
Normal rank

Latest revision as of 14:09, 20 June 2024

scientific article
Language Label Description Also known as
English
Mathematical aspects of outer-product asynchronous content-addressable memories
scientific article

    Statements

    Mathematical aspects of outer-product asynchronous content-addressable memories (English)
    0 references
    0 references
    1990
    0 references
    The paper summarizes and extends some important results on the dynamics (stability, attraction domains etc.) of several models of Hopfield networks. Canonical content addressable memories on n-cube (n power of 2) are introduced, in which there exist 2n vertices with mutual distances at least n/2, and their fixed points and attraction domains are more closely examined. (Rather pessimistic) results on relationships between memory capacity m/n (m-number of vectors to be stored) and \(\rho\) (fraction of initially incorrect bits in an input pattern) are formally derived.
    0 references
    0 references
    neural networks
    0 references
    stability
    0 references
    models of Hopfield networks
    0 references
    Canonical content addressable memories
    0 references
    fixed points
    0 references
    attraction domains
    0 references
    memory capacity
    0 references

    Identifiers